A significant body of preclinical and clinical work demonstrates the involvement of the serotonin system in alcoholism. Recent publication of two successful clinical trials of serotonin medications holds significant promise for the field. One study demonstrated a significant effect of the SSRI sertraline in late-onset (type A) alcoholics, while the other demonstrated efficacy of the 5-HT3-antagonist ondansetron in early-onset alcoholics. No other predictors of treatment response have been developed in alcoholism. Neuroimaging, particularly positron emission tomography (PET) scanning, offers a powerful tool to identify specific brain regions that may underlie or be associated with alcoholism. Neuroendocrine, physiological, subjective and recently [18 F] fluoro-deoxy-glucose (FDG) PET scanning responses to the broad spectrum serotonin agonist m-CPP have all been used to examine the serotonin system in alcoholism, and have previously been used to predict the clinical response to serotonin medications in other psychiatric disorders. FDG PET scanning has also recently begun to offer us an understanding of the regional pharmacotherapeutic treatment response to a serotonergic medication in depression, and in other disorders, but has yet to do so in alcoholism. We propose a clinical trial of a serotonergic medication in alcoholism, with the addition of a predictive serotonin neuroimaging and neuroendocrine probe before the study, and a follow up imaging scan at the end of the study. We will take 100 recently abstinent alcoholics, prescribe 200mg of sertraline together with weekly cognitive behavioral psychotherapy, and follow them for a 12-week period. We will perform an m-CPP FDG PET scan and a placebo FDG PET scan on completion of withdrawal prior to the trial, and an FDG PET scan on completion of the trial. Our aims are: firstly, to study changes in regional cerebral metabolism measured by FDG-PET scan induced by serotonin probe m-CPP relative to placebo FDG-PET scans in a group of alcoholics, and compare them with a group of healthy controls; secondly, to correlate the m-CPP induced changes in brain metabolism with the clinical response in a treatment trial of sertraline in the group of alcoholics over a 12-week period; and lastly, to assess the effect of sertraline on changes in placebo FDG- PET scans' regional metabolism, and to correlate changes in brain metabolism with treatment outcome. We hypothesize there will be group of clinical responders and non-responders in response to sertraline, and that the degree of serotonergic response, cerebral regional metabolic, hormonal, physiological or subjective, to the m-CPP challenge prior to the trial will predict the treatment response to sertraline in the trial. We would also hypothesize normalization of the FDG-PET scans in the treatment responsive sertraline group. This study should help identify a group of treatment responders to sertraline, examine whether that treatment response can be predicted using neuroimaging and neuroendocrine techniques, and identify what brain regions are implicated at baseline and in response to treatment.